To advance the multi-dimensional assessment and treatment of major depression, the broad innovations of this proposal are threefold. First, we will use functional magnetic resonance imaging (fMRI), and a within-subjects design with large sample sizes to identify neurobehavioral measures of mood, social functioning, and reward response in depression. Second, we will evaluate changes in these measures following cognitive behavioral therapy and assess the utility of these measures for predicting domain-specific treatment outcomes. Third, we will use these measures to develop fMRI-guided neurofeedback based on mood-derived brain states and test the capacity of the neurofeedback to elicit mood improvement. The overarching goal of this project is to identify quantitative neurobehavioral endophenotypes that may be used as multi-dimensional metrics in the assessment and treatment of major depressive disorder. The accompanying hypothesis is that anhedonia, social impairment, and negative mood in major depression may be mapped onto neural activity evoked during reward-, social functioning-, and emotion-based tasks, respectively. We believe that these measures may be used both to tailor individuals' treatment regimens and to develop novel personalized neurobehavioral therapies for depression. The specific aims are below: Aim 1: Identify and refine neurobehavioral measures of anhedonia, social functioning, and negative mood in individuals with major depressive disorder. We will use three distinct tasks and fMRI to quantify neurobehavioral responses to emotional stimuli, reward, and social functioning, in a within-subjects design in individuals with depression. We will test the hypothesis that aspects of major depression may be measured on a quantitative neurobehavioral spectrum of anomalies in these three domains. Aim 2: Identify domain-specific changes and predictors of treatment response to cognitive behavioral therapy. We will scan individuals with major depressive disorder before and after 12 weeks of empirically-supported cognitive behavioral therapy (CBASP) to identify a) domain-specific neurobehavioral predictors of treatment response and b) neurobehavioral concomitants of symptom changes elicited with CBASP. This Aim will test the hypothesis that the reward-, social-functioning, and mood- based neural responses identified in Aim 1 will serve differential predictive utility for domain-specific changes. Aim 3: Assess subjects' ability to modulate a neurofeedback interface developed from mood states and evaluate mood changes following neurofeedback. Subjects will receive neurofeedback derived from neural responses during positive or negative mood provocation. This aim will test the hypotheses that a) individuals can learn to modulate distributed brain patterns associated with positive and negative emotion, respectively, and that b) through neurofeedback, depressed individuals will achieve mood improvements.